docs(multiplayer): split spec into MULTIPLAYER_SPEC.md, trim task doc to outstanding work

multiplayer-next.md was a 1662-line mix of standing architecture spec
and task-completion tracking, most of which was dense per-task DONE
evidence for finished Phases 0-6. Split it:

- MULTIPLAYER_SPEC.md (new): the locked architecture decisions, wire
  format, server-side input handling, prediction/reconciliation,
  latency/frame-rate budget, and match lifecycle state machine -
  standing design reference, not task-tracked.
- multiplayer-next.md (trimmed 1662 -> ~370 lines): only outstanding
  work remains - §0 status, §7 Phase 7/8 task tables condensed to
  "what's left" per task, §8-11 reference material (refactoring notes,
  gotchas, testing, flagged items). Phases 0-6 collapsed to a pointer
  at git history instead of ~500 lines of DONE evidence.

Also:
- Repointed every `multiplayer-next.md §N` code comment (N 1-6) across
  Game/scripts, Game/tools and Game/tests to MULTIPLAYER_SPEC.md, since
  those sections moved. Task-number references (`task N.N`, §7-11)
  correctly still point at multiplayer-next.md.
- Updated CLAUDE.md's doc index and docs/TECH_STACK.md's spec-section
  citations to match.
- TODO.md: added a "what's left to actually finish multiplayer
  (human-actionable)" checklist pulled from multiplayer-next.md §0 and
  docs/MATCHMAKING.md - things that need a person (hardware, a design
  decision, a Steam App ID, hands on a controller), not more agent code.

Game/scripts/input_jitter_buffer.gd

@@ -1,7 +1,7 @@
 class_name InputJitterBuffer
 extends RefCounted
 
-# Per-player server-side input state (multiplayer-next.md §3, task 3.2).
+# Per-player server-side input state (MULTIPLAYER_SPEC.md §3; multiplayer-next.md task 3.2).
 # Deliberately a standalone RefCounted with no scene/RPC dependency — same
 # reason net_codec.gd and net_interpolator.gd are pure classes — so task
 # 3.5's unit tests can drive it with scripted arrival traces with no live
@@ -18,7 +18,7 @@ extends RefCounted
 # class's, since only the caller knows the current server tick.
 
 const RING_SIZE := 32
-# 500ms at 60Hz (multiplayer-next.md §3.2's own numbers) — a duration, not a
+# 500ms at 60Hz (MULTIPLAYER_SPEC.md §3.2's own numbers) — a duration, not a
 # tick-rate-derived constant, so left as a literal rather than pulling in
 # SimConstants for one number.
 const STARVE_ZERO_TICKS := 30

Game/scripts/input_lead_controller.gd

@@ -1,7 +1,7 @@
 class_name InputLeadController
 extends RefCounted
 
-# Client-owned input_lead control loop (multiplayer-next.md §3.3, task 3.3).
+# Client-owned input_lead control loop (MULTIPLAYER_SPEC.md §3.3; multiplayer-next.md task 3.3).
 # Standalone RefCounted, same reason as input_jitter_buffer.gd — scene-free
 # so it's directly unit-testable against scripted depth traces.
 #

Game/scripts/local_prediction_history.gd

@@ -3,7 +3,7 @@ extends RefCounted
 
 const NetBodyState = preload("res://scripts/net_body_state.gd")
 
-# Client-owned local-ship prediction history (multiplayer-next.md §4.3).
+# Client-owned local-ship prediction history (MULTIPLAYER_SPEC.md §4.3).
 # This is deliberately independent of NetworkedMatch and the scene tree so
 # sequence/ring behaviour can be tested from scripted traces. Each entry is
 # tagged with its full sequence number: an old value in a wrapped slot is

Game/scripts/match_net.gd

@@ -1,7 +1,7 @@
 extends Node
 
 # Autoload (project.godot [autoload] MatchNet). Handshake + roster layer on
-# top of NetworkManager's raw transport (§2.5, §1.3 of multiplayer-next.md).
+# top of NetworkManager's raw transport (§2.5, §1.3 of MULTIPLAYER_SPEC.md).
 # hello/welcome, strict protocol_version and physics_ticks_per_second
 # gating, player_joined/player_left, and — since lobby.tscn (task 1.5) needs
 # somewhere durable to keep it across the lobby→match scene transition —

Game/scripts/match_sim.gd

@@ -41,7 +41,7 @@ signal match_bootstrap_received(state: int, at_tick: int, score: Dictionary, end
 # §6.3 task 5.8: a spectator has been given a vacated slot at a kickoff.
 signal slot_assigned_received(peer_id: int, slot_index: int)
 
-# Input validation (multiplayer-next.md §3.1 steps 2-3, task 3.4). Deliberately
+# Input validation (MULTIPLAYER_SPEC.md §3.1 steps 2-3; multiplayer-next.md task 3.4). Deliberately
 # lives here rather than in NetworkedMatch: framing/rate abuse is a protocol-
 # level concern independent of any particular match's roster/slot state, and
 # this autoload already owns the RPC that receives the raw bytes.

Game/scripts/match_state.gd

@@ -1,6 +1,6 @@
 class_name MatchState
 
-# Match lifecycle states (multiplayer-next.md §6.1, task 5.1).
+# Match lifecycle states (MULTIPLAYER_SPEC.md §6.1; multiplayer-next.md task 5.1).
 #
 # Pure data + a transition table, deliberately with no scene, RPC or
 # NetworkedMatch dependency — same reason net_codec.gd and

Game/scripts/net_body_state.gd

@@ -1,6 +1,6 @@
 extends RefCounted
 
-# Plain data holder for one body's snapshot state (§2.4 of multiplayer-next.md).
+# Plain data holder for one body's snapshot state (§2.4 of MULTIPLAYER_SPEC.md).
 # Deliberately not Ship/Ball themselves, and deliberately not a scene-tree
 # node — NetCodec's pack/unpack must stay callable from pure-function tests
 # with no live scene. Phase 2's snapshot writer fills one of these per body

Game/scripts/net_codec.gd

@@ -1,7 +1,7 @@
 class_name NetCodec
 
 # Wire-format constants, quantisers, and pack/unpack for the two hot-path
-# packets (§2 of multiplayer-next.md). Pure functions only — no networking,
+# packets (§2 of MULTIPLAYER_SPEC.md). Pure functions only — no networking,
 # no autoload state — so they're testable head-on by tests/test_runner.tscn
 # without a live connection.
 #
@@ -48,7 +48,7 @@ const BODY_FLAG_STALLED := 1 << 5
 const BODY_FLAG_QUAT_W_SIGN := 1 << 6
 
 # --- Quantisation ranges (§2.4 — derived from arena/gameplay constants, not
-# restated prose; see multiplayer-next.md for the ArenaBoundary/Ship/Ball
+# restated prose; see MULTIPLAYER_SPEC.md for the ArenaBoundary/Ship/Ball
 # constants these are sized against) ---
 const POS_RANGE := 64.0        # metres, ±
 const VEL_RANGE := 64.0        # m/s, ±

Game/scripts/net_interpolator.gd

@@ -3,7 +3,7 @@ extends RefCounted
 
 # Buffers recent snapshot samples for ONE remote body and produces
 # interpolated states at any requested (possibly fractional) server tick —
-# used twice per body (multiplayer-next.md §4.1/§4.6, "dual-time remote
+# used twice per body (MULTIPLAYER_SPEC.md §4.1/§4.6, "dual-time remote
 # entities"): once at the present-time estimate for the collider, once
 # further back at present-minus-INTERP_DELAY for $Visual.
 #

Game/scripts/net_ship_predictor.gd

@@ -1,6 +1,6 @@
 extends RefCounted
 
-# Local-ship reconciliation policy (multiplayer-next.md §4.4).  Kept out of
+# Local-ship reconciliation policy (MULTIPLAYER_SPEC.md §4.4).  Kept out of
 # NetworkedMatch so the decision table is pure-testable; the imperative half
 # only writes Ship's existing Jolt-safe queued correction hooks.
 

Game/scripts/network_manager.gd

@@ -3,7 +3,7 @@ extends Node
 # Autoload (project.godot [autoload] NetworkManager). Owns transport-neutral
 # hosting, joining, shutdown, and connection-state signals. Lives
 # at a fixed autoload path so RPC NodePaths never depend on which scene is
-# loaded (§1.3 of multiplayer-next.md's derived decisions).
+# loaded (§1.3 of MULTIPLAYER_SPEC.md's derived decisions).
 #
 # server_relay = false is set the moment a peer exists: the default `true`
 # lets any client rpc() any other client *through the server*, which this

Game/scripts/ship.gd

@@ -145,7 +145,7 @@ func queue_teleport_with_velocity(to: Transform3D, new_linear_velocity: Vector3,
 	_has_pending_teleport = true
 
 
-# --- Netcode correction hooks (Phase 4; see multiplayer-next.md §4.4) ---
+# --- Netcode correction hooks (Phase 4; see MULTIPLAYER_SPEC.md §4.4) ---
 # Both stay zero until Phase 4 wires a reconciliation pass in, so the guarded
 # hook in _integrate_forces below is a no-op today.
 # Velocity delta from a soft correction, consumed once then cleared —

Game/scripts/sim_constants.gd

@@ -4,7 +4,7 @@ class_name SimConstants
 # constant derived from "60 Hz" (Ship._tick_scaled's decay reference,
 # reaction_ticks' export range, TrainingMode.TICKS_PER_SIM_SECOND) reads this
 # instead of restating the literal, so changing it changes every derived
-# constant coherently — see multiplayer-next.md §5.6 on why a future 120 Hz
+# constant coherently — see MULTIPLAYER_SPEC.md §5.6 on why a future 120 Hz
 # simulation needs to be a config change plus a retrain, not a protocol
 # rewrite hunting down bare 60s.
 #

Game/scripts/video_settings.gd

@@ -24,7 +24,7 @@ extends Node
 # independently of stretch mode, since it scales the 3D viewport's own internal
 # resolution before this blit rather than the window itself. Task 0.15b also
 # found an unexplained ~6% non-uniform width scaling on this project's one
-# tested (Mac/Retina) machine — see multiplayer-next.md §5.5.1 — which needs
+# tested (Mac/Retina) machine — see MULTIPLAYER_SPEC.md §5.5.1 — which needs
 # understanding before stretch mode is touched, not blindly carrying into a
 # resolution-dependent change.
 #
@@ -49,7 +49,7 @@ const SETTINGS_PATH := "user://settings.cfg"
 # preset -> bundle applied to the individual fields below. CUSTOM has no
 # bundle: selecting it just stops future preset changes from overwriting
 # whatever the individual fields currently hold. Task 0.15b's measured
-# per-effect costs (multiplayer-next.md §5.5.1) were too noisy to rank these
+# per-effect costs (MULTIPLAYER_SPEC.md §5.5.1) were too noisy to rank these
 # against each other, so each rung is "meaningfully fewer full-screen passes
 # than the one above it" rather than a precisely tuned ladder.
 const PRESET_BUNDLES := {
@@ -252,7 +252,7 @@ func apply_fps_cap() -> void:
 # Called once by each arena's _ready() (and again on settings_changed, so an
 # already-loaded arena updates live) to fold the user's glow/brightness
 # preference into that arena's own baked Environment tuning, and to gate the
-# preset-controlled full-screen passes (§5.5 of multiplayer-next.md).
+# preset-controlled full-screen passes (§5.5 of MULTIPLAYER_SPEC.md).
 func apply_to_environment(env: Environment) -> void:
 	if env == null:
 		return